The following directories can hold rather large binary log files that can fill up /var if there was an error. The docs should explain what they are for and that they are safe to delete.
I would expect to find it on the admin maintain and tune page.
Relevant IRC chatter from #puppet.
12:51 < blkperl> grim_radical: discarded and localhost
12:51 < grim_radical> blkperl: if there's a bunch of stuff in discarded, then those are processing failures
12:52 < grim_radical> you can look in that directory and crack open the files; they should have a bit-for-bit copy of the input data, along with
every retry timestamp, and the full trace of each error encountered
12:52 < blkperl> grim_radical: so safe to delete?
12:53 < blkperl> looks like they are about an ssh-key input failure from a few days ago
12:53 < grim_radical> blkperl: yeah, you can nuke the stuff underneath discarded and it should be fine